Primary Analysis
From Table 1, seven models are presented analyzing changes in probability of sign up.
Table 1: Logistic Regression Output
(est1 stored)
(est2 stored)
(est3 stored)
(est4 stored)
(est5 stored)
(est6 stored)
(est7 stored)
Table 1: Logistic Regression on Contract Adoption
----------------------------------------------------------------------------------------------------------------------------------
> --
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
>
----------------------------------------------------------------------------------------------------------------------------------
> --
y_signup
>
5% savings rate 0 0 0 0 0 0 0
>
(.) (.) (.) (.) (.) (.) (.)
>
10% savings rate 0.123 0.124 0.111 0.117 0.118 0.157 0.157
>
(0.104) (0.104) (0.104) (0.105) (0.105) (0.109) (0.109)
>
20% savings rate 0.0934 0.0830 0.0771 0.0768 0.0777 0.117 0.120
>
(0.101) (0.101) (0.102) (0.102) (0.102) (0.107) (0.107)
>
01 High 0 0 0 0 0 0
>
(.) (.) (.) (.) (.) (.)
>
02 Moderate income 0.347* 0.304* 0.303* 0.308* 0.178 0.177
>
(0.140) (0.142) (0.142) (0.143) (0.149) (0.149)
>
03 Low income -0.222 -0.306 -0.231 -0.231 -0.356* -0.348*
>
(0.153) (0.158) (0.161) (0.161) (0.173) (0.174)
>
01 White 0 0 0 0 0
>
(.) (.) (.) (.) (.)
>
02 Black 0.0853 0.168 0.164 0.0908 0.0749
>
(0.165) (0.171) (0.171) (0.174) (0.175)
>
03 Hispanic 0.233 0.313 0.321 0.170 0.161
>
(0.184) (0.186) (0.186) (0.203) (0.205)
>
04 Asian -0.143 -0.0725 -0.0702 -0.0155 0.00504
>
(0.170) (0.173) (0.174) (0.186) (0.186)
>
05 Other POC 0.327 0.379 0.377 0.442* 0.444*
>
(0.193) (0.194) (0.194) (0.204) (0.204)
>
01 Homeowner 0 0 0 0
>
(.) (.) (.) (.)
>
02 Renter -0.315* -0.326** -0.0908 -0.0694
>
(0.125) (0.125) (0.132) (0.133)
>
03 Other 0 0 0 0
>
(.) (.) (.) (.)
>
10-page contract 0 0 0
>
(.) (.) (.)
>
20-page contract -0.0637 -0.0661 -0.0564
>
(0.0828) (0.0856) (0.0856)
>
1-year contract 0 0 0
>
(.) (.) (.)
>
25-year contract -0.222* -0.228* -0.225*
>
(0.0898) (0.0926) (0.0927)
>
Zero Cancellation ~e 0 0 0
>
(.) (.) (.)
>
$250 Cancellation ~e 0.0131 0.0385 0.0419
>
(0.0855) (0.0905) (0.0909)
>
01 Less familiar 0 0
>
(.) (.)
>
02 More familiar 1.471*** 1.479*
> **
(0.123) (0.123)
>
01 Less review 0
>
(.)
>
02 More review 0.255*
>
(0.116)
>
Constant 0.314*** 0.217 0.201 0.259 0.395* -0.385* -0.560*
> *
(0.0820) (0.124) (0.136) (0.139) (0.159) (0.172) (0.188)
>
----------------------------------------------------------------------------------------------------------------------------------
> --
Observations 2986 2986 2986 2976 2976 2976 2976
>
----------------------------------------------------------------------------------------------------------------------------------
> --
Standard errors in parentheses
* p<0.05, ** p<0.01, *** p<0.001
Table 2: Odds Ratios
(est1 stored)
(est2 stored)
(est3 stored)
(est4 stored)
(est5 stored)
(est6 stored)
(est7 stored)
Table 1: Logistic Regression on Contract Adoption
----------------------------------------------------------------------------------------------------------------------------------
> --
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
>
----------------------------------------------------------------------------------------------------------------------------------
> --
y_signup
>
5% savings rate 1 1 1 1 1 1 1
>
(.) (.) (.) (.) (.) (.) (.)
>
10% savings rate 1.131 1.132 1.118 1.124 1.125 1.170 1.170
>
(0.117) (0.117) (0.116) (0.118) (0.118) (0.127) (0.128)
>
20% savings rate 1.098 1.086 1.080 1.080 1.081 1.124 1.127
>
(0.111) (0.110) (0.110) (0.110) (0.110) (0.120) (0.121)
>
01 High 1 1 1 1 1 1
>
(.) (.) (.) (.) (.) (.)
>
02 Moderate income 1.414* 1.355* 1.354* 1.360* 1.195 1.193
>
(0.198) (0.193) (0.193) (0.194) (0.178) (0.178)
>
03 Low income 0.801 0.736 0.794 0.793 0.700* 0.706*
>
(0.123) (0.116) (0.128) (0.128) (0.121) (0.123)
>
01 White 1 1 1 1 1
>
(.) (.) (.) (.) (.)
>
02 Black 1.089 1.183 1.178 1.095 1.078
>
(0.180) (0.202) (0.201) (0.191) (0.189)
>
03 Hispanic 1.263 1.367 1.378 1.185 1.175
>
(0.232) (0.255) (0.256) (0.240) (0.241)
>
04 Asian 0.867 0.930 0.932 0.985 1.005
>
(0.147) (0.161) (0.162) (0.184) (0.187)
>
05 Other POC 1.386 1.461 1.458 1.555* 1.558*
>
(0.268) (0.283) (0.283) (0.317) (0.319)
>
01 Homeowner 1 1 1 1
>
(.) (.) (.) (.)
>
02 Renter 0.730* 0.722** 0.913 0.933
>
(0.0910) (0.0901) (0.121) (0.124)
>
03 Other 1 1 1 1
>
(.) (.) (.) (.)
>
10-page contract 1 1 1
>
(.) (.) (.)
>
20-page contract 0.938 0.936 0.945
>
(0.0777) (0.0802) (0.0809)
>
1-year contract 1 1 1
>
(.) (.) (.)
>
25-year contract 0.801* 0.796* 0.799*
>
(0.0720) (0.0737) (0.0740)
>
Zero Cancellation ~e 1 1 1
>
(.) (.) (.)
>
$250 Cancellation ~e 1.013 1.039 1.043
>
(0.0866) (0.0941) (0.0948)
>
01 Less familiar 1 1
>
(.) (.)
>
02 More familiar 4.355*** 4.388*
> **
(0.534) (0.539)
>
01 Less review 1
>
(.)
>
02 More review 1.291*
>
(0.149)
>
----------------------------------------------------------------------------------------------------------------------------------
> --
Observations 2986 2986 2986 2976 2976 2976 2976
>
----------------------------------------------------------------------------------------------------------------------------------
> --
Exponentiated coefficients; Standard errors in parentheses
* p<0.05, ** p<0.01, *** p<0.001
Controlling for all other variables, contract length was found to have a strong negative relationship on the probability of contract adoption. For example, holding all variables constant, from model 5 we see that the odds ratio for contract length of 25 years is 0.82, meaning the group with contracts of 25 years are 0.82 times as likely as the group with contracts of 1 year of signing up.
No statistically significant difference in contract adoption were observed when respondents are shown differing savings rates, suggesting other attributes are primary motivations instead of savings rate. Relative to White respondents, ‘Other POC’ were 1.55 times more likely to adopt the contract in model 6, controlling for all other demographic and contract attributes.
Moderate income individuals were more likely relative to high income individuals to accept the contract when controlling for race, contract length in both pages and years, and cancellation fees. However, by model 6, the difference is no longer statistically significant. Interestingly, in model 6 and 7, low income individuals,, were less likely to adopt the contract relative to high income and controlling for all variables